At the Snapdragon Tech Summit 2023 event held in Hawaii, Qualcomm announcedSnapdragon X EliteAndSnapdragon 8 Gen3The two main processor products emphasize the integration of self-generated artificial intelligence on the hardware side, symbolizing that the future development of mobile computing will be in the form of 5G combined with artificial intelligence, but with a greater emphasis on accelerated computing on the device side.
Compared to previous presentations, Qualcomm's introduction of its new mobile computing platform, while still emphasizing architectural design details and performance improvements, also took considerable time to explain the possibilities of on-device AI applications and the new user experience enabled by automatically generated AI.
Many of the application functions demonstrated this time were previously accessible only through an internet connection, such as digital assistant services enabled by cloud-based collaborative computing or image processing capabilities enabled by networked computing. However, by integrating AI technology into the device, these computing functions can be realized without relying on a network connection and collaborative computing, completing tasks faster and even ensuring that personal privacy data remains on the device without having to be uploaded to the cloud.
By enabling the device to execute large-scale natural language models and respond to various automatically generated artificial intelligence application services, including digital assistant services that can provide information opinions and help organize content, or speed up the processing of HDR images taken by the device camera, and even automatically generate image content outside the range of the original photo through deep learning methods, these applications can complete calculations on the device side without the need for an Internet connection, which means that upcoming mobile phones and PC devices will be able to integrate more artificial intelligence application functions, while also bringing a faster and more diverse operating experience.
The integration of self-generated artificial intelligence means that future mobile phones will better understand users' actual needs and can also serve as personal assistants to help provide more information needed at the moment, and even allow users to interact with their phones through natural language conversations (Note).
Note:Qualcomm currently only supports English, but this feature can be expanded later, or OEMs and ODMs can add more languages when using their processor products.
Qualcomm believes that the core configuration, architecture, performance, power consumption and other aspects of the processor design are still important, but future developments will focus more on integrating artificial intelligence technology, using automatically generated artificial intelligence to change the existing mobile phone user experience while also accelerating the overall operational response of the phone.
However, simply using device-side AI to speed up operational responses is still insufficient. Therefore, hybrid architecture computing will continue to be the mainstream in the future. The collaborative computing model combining device-side and cloud computing will continue to be the mainstream. However, edge computing on the device side will be greatly improved by automatically generated AI. The cloud computing model will only be combined when higher computing power resources are required or when a new version of the database needs to be downloaded. Otherwise, the automatically generated AI on the device side will most likely be used to complete the calculation first.
Therefore, if the basic usage patterns do not change much, the future experience of using mobile phones will significantly increase the proportion of device-side priority computing.
Related links:
• Snapdragon X Elite is Qualcomm's first PC-class processor built with NUVIA technology
